In 1785, a French mathematician published a book about voting that almost nobody read, yet it has been quietly explaining elections ever since.
The Marquis de Condorcet was the permanent secretary of the Académie des Sciences, and his immediate problem was surprisingly domestic. The academy could not agree on how to elect its own members. His colleague, Jean-Charles de Borda, had proposed a points system.
Condorcet worked through the arithmetic and discovered something quite unsettling. The candidate who received the most first-choice votes was not necessarily the candidate the assembly preferred overall.
One candidate could win by simply having the most first choices, while another could beat every other candidate when voters were asked to choose between them one at a time.
The second candidate became known as the Condorcet winner. He was not necessarily the person most people ranked first, but he was the person who could beat every other candidate in a one-to-one contest.
In the final decades of the Roman Republic, a similar structural dynamic played out. No single faction could permanently dominate the others.
Each major grouping including the populares, the optimates, and the provincial mercantile classes, could block its rivals far more easily than it could assemble a lasting majority of its own. The result was a political culture defined by successive vetoes. The figure who endured was rarely the most loved.
He was the one who lost least often when every alternative was measured against him in isolation. Continuity became the residual outcome of incompatible ambitions.
Today, Kenya is witnessing in real time what unsettled Condorcet.
Over the last couple of weeks, we have heard similar sentiments. Voters have said that if their man is not the preferred opposition presidential candidate, they will have no incentive to vote for an opposition candidate from another region, and might as well vote for Ruto or stay home.
What they are simply saying is, my candidate first, Ruto second, the other candidates after that. The phrasing changes. The ordering does not.
We tend to hear these statements as evidence of tribal loyalty. But there is something more interesting happening underneath.
Each community is describing not only its preferred candidate, but also its second choice. And in a presidential election, the second choice can matter as much as the first.
This is the Condorcet logic of Kenyan politics currently. Ruto doesn’t need to be anyone’s first choice. He needs only to be the least unacceptable alternative once the ethnic vetoes have been registered.
Continuity becomes the residual option of communities that have not secured their own elevation. If this hierarchy of preference holds, then Ruto beats each of them in any head-to-head contest.
And this is not the vote-splitting our political commentary keeps describing. Vote-splitting happens when several candidates divide the support of people who would otherwise prefer the same side, allowing another candidate to win with a plurality. That is a coordination failure, and hence, coordination can solve it.
Here something different is happening. There is a hierarchy of preference which if it holds, bringing the opposition candidates together will not necessarily change the outcome. It would simply confirm it.
Begs the question. Why is Ruto the second choice?
The uncomfortable truth is tribalism. However, it explains very little because it simply gives a name to the behaviour without explaining it. The more useful question should be: what is the voter who prefers Ruto, if their preferred principal is not the presidential candidate, trying to protect?
Let me unpack it for you. Kenyan presidential politics runs on a queue, which is a sequence in which each large community holds an expectation of its turn. ‘It is our turn’ is not an idle slogan. It is a claim on a schedule. And the defining property of a queue is that your place in it depends on what other people do.
So consider the price a Luhya voter pays to support a Kamba candidate. It is not one election, but the 10 years that candidate would occupy the top seat, and worse, it concedes a claim that had not previously been conceded. You have not merely lost the round. You have used your own vote to move yourself farther back in the queue.
The opposition have been telling us they will unite to oust Ruto. We often speak of political coalitions as though they are formed by adding votes. But coalitions are also formed by asking somebody else to wait.
The presidential election is therefore not simply a contest over who can assemble the largest coalition. It is a negotiation over whose ambition gets deferred and whose gets advanced. And the voter asks, if my man cannot win, why should I help your man win?
Now compare that with voting for the incumbent. His second term costs nothing in queue terms. If the precedent holds whereby each past president has served two terms, it means the cost in queue terms was already priced in 2022.
Therefore, supporting him again does not create a new 10-year wait for other communities and hence does not push another community farther back than it already expected to be. It starts no new clock.
This explains why anti-incumbency can be entirely genuine but still fail to convert. Voters may be genuinely angry but acting on that rage costs them more than enduring the status quo.
They may dislike the incumbent and still prefer him to the alternative available to them. The result is a curious form of political inertia where dissatisfaction grows, but it does not translate into a vote for change because the perceived cost of change remains higher.
You can watch the queue being traded publicly. Within Linda Mwananchi, Babu Owino has reportedly said that he would want any winner from the formation to serve a single term and hand over to him in 2032. This is a man bargaining over sequence.
So what kind of politics emerges when the most viable national candidate is the one who can defeat regional kingpins in pairwise contests?
The first effect is that the opposition is forced into a difficult trade-off. To maximise strength inside one community, a leader must emphasise ethnic identity and grievance.
Yet the more successfully he does so, the harder it becomes to persuade other communities that he is a national option rather than someone else’s champion. The same quality that makes him a Condorcet threat inside his base makes him a Condorcet loser outside it.
In other words, the loyalty he commands among his core supporters is matched by the resistance he provokes among voters who have already decided that he is unacceptable.
Coalition-building therefore becomes an exercise in managing incompatible first preferences rather than constructing a shared programme.
The second effect is that the incumbent is handed a structural cushion. He does not have to inspire broad enthusiasm.
He only has to remain preferable to each of the main ethnic alternatives in the eyes of everyone else. Although performance still matters, it is filtered through this comparative frame.
A voter may be disappointed with the current administration and still conclude that the available alternatives are, from their communal point of view, worse.
The third effect is deeper. When political choice is organised as a series of ethnic pairwise contests, the citizen is invited to think first as a member of a community and only second as a participant in a national judgement about competence.
The question who would govern best is displaced by the question, whose absence is least tolerable? The Condorcet winner in such a system is not necessarily the most capable or popular. It is the candidate who best survives the vetoes.
This is not an argument that ethnicity should be ignored, because it is a real and historically rooted force. The danger lies in allowing it to become the primary sorting mechanism of political viability.
Once every major community announces that its support is conditional on the elevation of a particular individual, the space for a politics of ideas and performance narrows.
The incumbent benefits from the fragmentation, the opposition struggles to escape it and the country is left with a politics that is highly responsive to ethnic anxiety and relatively unresponsive to anything else.
Condorcet exposed a vulnerability inherent in multi-candidate systems, which was that intense passion rarely translates into broad-based stability.
A candidate in a coalition can command a passionate 35 per cent of voters who rank him first above all others, yet remain deeply reviled by the remaining 65 per cent.
If that 65 per cent would choose almost anyone else in a head-to-head contest, then the candidate with the most first-place votes may, in fact, be the most vulnerable candidate on the ballot.
This mathematical reality provides a useful lens for understanding Ruto’s strategic calculations. Could it be that the formation of the broad-based government was more than a short-term tactical manoeuvre designed to navigate short-term political crises?
Seen through the framework of social choice theory, was it a calculated attempt to engineer a Condorcet equilibrium, where the governing coalition becomes the median preference of a divided national electorate? You be the judge.
Finally, my unsolicited advice is to Ruto. Do not take the Condorcet advantage for granted. If the logic holds, your strongest electoral asset may not be enthusiasm, but the reluctance of different communities to hand their political future to another.
This is a powerful cushion, but also a fragile one. Performance still matters because disappointment can eventually become greater than the fear of the alternative. You therefore have an unusual opportunity.
Rather than trying to become every community’s first choice, you need to remain the candidate that communities can live with when their first choice is not available. In Condorcet terms, that may be enough to win.
Of two evils, the lesser is always to be chosen – Aristotle
